github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
nine So, the morale of the story is: beware of the closure! Failures can become serialized, even if they are just the result of an indirect lookup and the result is just used in 2 lines of code if those lines create a lexical variable and further down the function a closure is created. 06:58
Even if the closed over code doesn't actually use that variable!
And of course, this being Perl 6, even if statements can create closures... 06:59
Kaiepi can M#1166 be merged today? i was hoping to get started on sync udp/unix/raw socket support today 10:19
synopsebot M#1166 [open]: github.com/MoarVM/MoarVM/pull/1166 [IP6NS Grant] Improve hostname resolution
lizmat Kaiepi: it seems to have conflicts ? 11:51
in src/io/syncsocket.c
Kaiepi ah, i'll fix those in a sec 11:59
Kaiepi lizmat, fixed 12:15
lizmat the leak issue is also fixed? colabti.org/irclogger/irclogger_lo...9-09-06#l1 12:16
Kaiepi yes 12:19
Geth_ MoarVM/master: 8 commits pushed by (Ben Davies)++, (Elizabeth Mattijsen)++
lizmat Kaiepi: have fun!
Kaiepi sweet, thanks 12:21
brrt \o 12:40
nwc10 o/ 12:57
brrt ohai nwc10 13:04
Guest3346 brrt, should we remind you of sp_bind or whatever the name was? 13:37
brrt yes, that's a decent plan 13:38
it's on the back of my mind
it's just very annoying to debug
Guest3346 you mentioned something like that some time ago 13:39
are the bisect tools more or less unusable in this case? 13:40
brrt I'm not quite sure why I haven't been able to pin it down yet 13:59
Guest3346 perhaps you just need a few more minutes of debugging before the solution presents itself 14:07
brrt hehe 14:17
I think you may underestimate the difficulty of debugging JIT problems
more like hours
Guest3346 I might have underestimated the the problem a tiny bit :-) 14:24
lizmat And another Perl 6 Weekly hits the Net: p6weekly.wordpress.com/2019/10/07/...syntaxing/
brrt it's probably a simple problem, but a JIT has many moving parts, most of them not directly interfacing with the user, so hard to trigger and test 14:27
lizmat++ 14:28
hankache Hello * 19:58
I found a weird issue while building Rakudo on WIndows 10 19:59
I tried three sets of settings:
1) perl Configure.pl --backends=moar --gen-moar --moar-option="--cc=cl" --moar-option="--ld=link" --moar-option="--make=nmake" --relocatable
2) perl Configure.pl --backends=moar --gen-moar --moar-option="--cc=cl" --moar-option="--make=nmake" --relocatable 20:00
3) perl Configure.pl --backends=moar --gen-moar --moar-option="--cc=gcc" --moar-option="--make=gmake" --relocatable
brrt and what happened? 20:01
hankache All of them build correctly and install most modules I tested except for LWP::Simple
using option 1 the tests of LWP::Simple fails 20:02
PS: Windows 10
The culprit seems to be --moar-option="--ld=link" 20:04
This is the failing test: github.com/perl6/perl6-lwp-simple/...tore.t#L38 20:05
MasterDuke++ proposed I raise the issue here instead of #perl6-dev 20:06
Am I missing something? 20:13
dogbert17 valgrind seems to be upset with t/spec/S17-procasync/encoding.t 21:28
==17964== Invalid read of size 4 21:29
==17964== at 0x5155F4A: MVM_string_utf8_decodestream (utf8.c:453)

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!